1
\$\begingroup\$

Below shows two types of CM choke winding:

enter image description here

Image source: Common Mode Chokes - Basic Knowledge | Würth Elektronik

Is there a difference between the two in application? They are both used in literature, yet I haven't seen a single sentence about the difference.

\$\endgroup\$
3
  • 2
    \$\begingroup\$ Just looking at them: the one on the left will have less parasitic inductance (i.e., differential mode), while the one on the right will have less interwinding capacitance. \$\endgroup\$
    – Evan
    Dec 7, 2018 at 18:56
  • \$\begingroup\$ Does that mean the one on the riight will attenuate differential mode signal moe?. \$\endgroup\$
    – floppy380
    Dec 7, 2018 at 19:24
  • \$\begingroup\$ This topic is crazy there is no good explanation on net \$\endgroup\$
    – floppy380
    Dec 7, 2018 at 19:24

4 Answers 4

2
\$\begingroup\$

The second type can be made with a large spacing between the wires so that it's suitable for mains differential voltage. The one shown below has a 1500VAC test voltage. It is suited for high voltage power filtering.

enter image description here

The first type can be manufactured with controlled differential impedance, so it is suited to high frequency signal filtering.

enter image description here

\$\endgroup\$
3
  • \$\begingroup\$ If the choke is used right at the output of a DC power supply to supress 1Meg to 5MegHz CM interference,do both type of winding work? \$\endgroup\$
    – floppy380
    Dec 7, 2018 at 19:22
  • \$\begingroup\$ They'll both work, given proper core material for the purpose, The left hand one might allow a somewhat smaller core for the same differential current. \$\endgroup\$ Dec 7, 2018 at 19:45
  • \$\begingroup\$ Thank you Im very new to the subject please have a look at my other question if you have time: electronics.stackexchange.com/questions/411067/… \$\endgroup\$
    – floppy380
    Dec 7, 2018 at 23:52
3
\$\begingroup\$

The choke on the left maintains a somewhat constant spacing between wires thus maintaining a similar characteristic impedance and thus is better suited for RF applications than the choke on the right which makes no effort to maintain a constant characteristic impedance throughout the choking function. For RF applications, a constant stable characteristic impedance is a desirable feature for chokes.

\$\endgroup\$
1
\$\begingroup\$

Checking on this question some years later, it's not clear what trouble you had finding resources; but it does often happen that, lacking the correct keywords, little relevant information turns up. In that spirit, you may find these resources of interest.

An image search turns up these references:
How Common Mode Chokes Work | EMI Analyst
Common Mode Chokes - Basic Knowledge | Würth Elektronik

You may also find the transformer 2nd-order equivalent circuit of interest:

enter image description here
Source: my website, https://www.seventransistorlabs.com/Images/XfmrEquiv.png

Of course as a CMC, we turn this sideways, but either way, we end up with a network of capacitances and inductances, which can be measured at various frequencies, and with open-circuit, short-circuit, or terminated (characteristic impedance) conditions. These conditions allow us to isolate one or another element and measure it fairly directly, or in combinations that are easy to identify (such as resonant circuits).

You may also find the Wikipedia article on Leakage inductance of interest. The modeling, and measurement, of LLp + LLs, and the relation to T1 magnetizing inductance, is most useful.

Physically speaking, anywhere we have two lengthy conductors spaced apart, we have some inductance and capacitance between them, and a transmission line of sorts -- with characteristic impedance and electrical length (that is, a time delay, corresponding to the physical length of the conductors divided by speed of light in the surrounding medium). More specifically, the inductance and capacitance are either infinitesimal elements of that transmission line, or the low-frequency equivalent totals.

Since we're mostly concerned about conducted EMI frequencies for CMCs, the LF equivalent applies, and we don't have to worry too much about transmission-line effects. TL effects are often relevant for data-line type chokes, where the impedance indeed might be controlled to maintain signal quality.

Returning to the image in question, we might use a close-spaced pair (often twisted) where a consistent TL impedance (that is, the spacing between wires is consistent) is desirable, or where LL shall be kept low. Conversely, we might wind the wires in separate sectors to maximize distance between them, maximizing leakage and isolation voltage (separation distance).

Note that, because core materials are generally at least moderately conductive, the capacitance between windings is not actually reduced all that much, comparing between the two cases. When a high SRF (self-resonant frequency) is required, a smaller core with fewer turns is generally necessary.

Incidentally, you do occasionally find line-type CMCs of the first style, where one or both wires are made from TIW (triple-insulated wire) to achieve the necessary isolation voltage rating. They can have higher inductance/impedance or more compact size compared to the other style. The downside is, the low leakage inductance typically doesn't help with differential-mode filtering, and a separate DM choke may still be required.

\$\endgroup\$
0
\$\begingroup\$

Look closely at the windings on the 2 chokes. Left one is wound for common mode cancellation. The right is wound for differential mode. The header statement is not true. They are not both CM type chokes

\$\endgroup\$
1
  • \$\begingroup\$ I mean, you could wire it that way, it's whatever, but the two wires entering on the same side of the core (in both senses of "same side") is very typical for a CMC. \$\endgroup\$ Jan 15 at 19:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.